Agents with Attitude: Exploring Coombs Unfolding Technique with Agent-Based Models

This snapshot is intended as an illustration of some ways that computational modeling can help one engage with and extend pragmatically complex applied mathematical problems. Specifically, I explore Coombs unfolding technique (1950), a mathematical method used in psychometrics to analyze individual and collective preferences toward some topic of interest. Coombs unfolding works well when used on individuals, but becomes difficult if one wishes to combine individual data to determine the attitudes and preferences of an entire population. In this paper, I briefly introduce Coombs unfolding technique, and model the technique using simple agent-based models (ABMs) developed in the NetLogo (Wilensky 1999) modeling environment. I then attempt to address some of the difficulties one may encounter when applying Coombs unfolding to a population by using these ABMs as a starting point. Finally, I discuss the affordances and disadvantages of using computational modeling as a starting point for thinking about applied mathematical problems.